Optimization of waste paper´s enzymatic deinking processes based on neural network and particle swarm optimization

The optimal technological parameters of a waster paper´s enzymatic deinking process with strong coupling, nonlinear and large time delay are difficult to achieve. BP neural network and improved particle swarm optimization (PSO) were applied to optimize enzymatic deinking process of waste paper. The theory and process were described. Enzymes dosage, temperature and pH were used as inputs of the network, a BP neural network model of the effective residual ink concentration (ERIC) and brightness of the pulp was established. The model had higher prediction precision compared with traditional regression model. The PSO was used to obtain the optimal conditions of deinking process with the lowest ERIC and highest brightness of the pulp. Experiments´ results proved the method was an excellent tool for optimization of enzymatic deinking process.